CN116029329A - Anxiety mileage value prediction method, anxiety mileage value prediction device, anxiety mileage value prediction system and storage medium - Google Patents

Anxiety mileage value prediction method, anxiety mileage value prediction device, anxiety mileage value prediction system and storage medium Download PDF

Info

Publication number
CN116029329A
CN116029329A CN202310115798.3A CN202310115798A CN116029329A CN 116029329 A CN116029329 A CN 116029329A CN 202310115798 A CN202310115798 A CN 202310115798A CN 116029329 A CN116029329 A CN 116029329A
Authority
CN
China
Prior art keywords
anxiety
mileage
original
value
positions
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310115798.3A
Other languages
Chinese (zh)
Inventor
邬少飞
刘晗
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan Institute of Technology
Original Assignee
Wuhan Institute of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan Institute of Technology filed Critical Wuhan Institute of Technology
Priority to CN202310115798.3A priority Critical patent/CN116029329A/en
Publication of CN116029329A publication Critical patent/CN116029329A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention provides a method, a device, a system and a storage medium for predicting anxiety mileage values, which belong to the field of data prediction, wherein the method comprises the following steps: s1: carrying out data cleaning on the original electric vehicle running data to obtain cleaned running data; s2: screening and analyzing the cleaned driving data to obtain a mileage anxiety data set; s3: dividing the mileage anxiety data set to obtain a mileage anxiety training set and a mileage anxiety testing set; s4: training the training model according to the mileage anxiety training set to obtain an original prediction model. Compared with the existing anxiety mileage value prediction method, the method overcomes the defects of large workload, easy sinking into local optimum and poor prediction and identification of the electric automobile driver of the traditional anxiety mileage value prediction method, can accurately predict the anxiety mileage of the electric automobile driver, and solves the problem of insufficient prediction of the anxiety mileage value of the existing electric automobile driver.

Description

Anxiety mileage value prediction method, anxiety mileage value prediction device, anxiety mileage value prediction system and storage medium
Technical Field
The invention mainly relates to the technical field of data prediction, in particular to a method, a device and a system for predicting an anxiety mileage value and a storage medium.
Background
At present, there are some researches on mileage anxiety at home and abroad. Regarding the cause of mileage anxiety, franke considers that researchers lack scientific knowledge on mileage experience of electric automobile travelers, actual psychological anxiety degree of the travelers cannot be measured correctly, and he proposes a concept of 'comfortable mileage' to solve the perception problem of mileage experience. Some students have conducted related searches from the direction of recognition range anxiety. Li Zonghua the influence of the mileage anxiety of the pure electric vehicle user on the user use behavior is researched, and a mileage anxiety degree judgment model based on the Internet of vehicles big data user use behavior is provided. There are also scholars who have studied the main factors affecting the range. According to the study, in addition to the current SOC (referring to the proportion of the current electric quantity to the rated electric quantity), the electric vehicle driver more hopefully predicts the remaining mileage accurately than the increase in the battery capacity. Since mileage anxiety is commonly found in electric vehicle users, research on mileage anxiety is significant. Through test investigation, rauh Nadine et al obtain that the mileage anxiety degree has certain relevance with the driving experience enrichment degree of the user.
There are also scholars who have studied the main factors affecting the range. According to the study, in addition to the current SOC (referring to the proportion of the current electric quantity to the rated electric quantity), the electric vehicle driver more hopefully predicts the remaining mileage accurately than the increase in the battery capacity. Air conditioning is one of the largest accessory energy consumption in electric vehicles, so whether the air conditioner is on and the power level have an important influence on the prediction of driving range. Also, the impact of the battery is of paramount importance. At a lower temperature, the charge and discharge capacity of the battery is lower than the normal temperature, the internal resistance is increased in a nonlinear manner, and the battery consumes more energy, so that the driving range of the electric automobile is limited. Xie Chi et al consider traffic network equalization problems based on mileage anxiety factors. Wang Tao et al consider a charging station site selection model based on mileage anxiety.
In summary, it can be seen that each electric vehicle user has more or less mileage anxiety, and both electric vehicle developers and electric vehicle users are concerned about mileage anxiety, so that the electric vehicle users have great significance in researching mileage anxiety. Throughout domestic and foreign researches, researchers have more research applications based on mileage anxiety, and the users are better served by using mileage anxiety, while the electric automobile driver is researched based on history data of the electric automobile, and the prediction of anxiety mileage is relatively less.
Disclosure of Invention
The invention aims to solve the technical problem of providing a method, a device, a system and a storage medium for predicting anxiety mileage values aiming at the defects of the prior art.
The technical scheme for solving the technical problems is as follows: a method for predicting anxiety mileage values, comprising the steps of:
s1: importing a plurality of original electric vehicle running data, and cleaning all the original electric vehicle running data to obtain a plurality of cleaned running data;
s2: screening and analyzing all the cleaned driving data to obtain a mileage anxiety data set;
s3: dividing the mileage anxiety data set to obtain a mileage anxiety training set and a mileage anxiety testing set;
s4: building a training model, and training the training model according to the mileage anxiety training set to obtain an original prediction model;
s5: testing the original prediction model according to the mileage anxiety testing set to obtain a target prediction model;
s6: and importing the running data of the electric automobile to be predicted, and predicting the running data of the electric automobile to be predicted through the target prediction model to obtain a prediction result of the anxiety mileage value.
The other technical scheme for solving the technical problems is as follows: an anxiety mileage value prediction device, comprising:
the data cleaning module is used for importing a plurality of original electric vehicle running data, and cleaning the data of all the original electric vehicle running data to obtain a plurality of cleaned running data;
the screening analysis module is used for screening and analyzing all the cleaned driving data to obtain a mileage anxiety data set;
the division module is used for dividing the mileage anxiety data set to obtain a mileage anxiety training set and a mileage anxiety testing set;
the model training module is used for constructing a training model, and training the training model according to the mileage anxiety training set to obtain an original prediction model;
the model test module is used for testing the original prediction model according to the mileage anxiety test set to obtain a target prediction model;
the prediction result obtaining module is used for importing the running data of the electric automobile to be predicted, and predicting the running data of the electric automobile to be predicted through the target prediction model to obtain a prediction result of the anxiety mileage value.
Based on the anxiety mileage value prediction method, the invention further provides an anxiety mileage value prediction system.
The other technical scheme for solving the technical problems is as follows: an anxiety distance value prediction system comprising a memory, a processor and a computer program stored in the memory and executable on the processor, which when executed by the processor implements the anxiety distance value prediction method as described above.
Based on the anxiety mileage value prediction method, the invention also provides a computer readable storage medium.
The other technical scheme for solving the technical problems is as follows: a computer readable storage medium storing a computer program which, when executed by a processor, implements the anxiety mileage value prediction method as described above.
The beneficial effects of the invention are as follows: the method comprises the steps of cleaning data of original electric vehicle driving data to obtain cleaned driving data, screening and analyzing the cleaned driving data to obtain a mileage anxiety data set, dividing the mileage anxiety data set to obtain a mileage anxiety training set and a mileage anxiety testing set, training a training model according to the mileage anxiety training set to obtain an original prediction model, testing the original prediction model according to the mileage anxiety testing set to obtain a target prediction model, and predicting the driving data of the electric vehicle to be predicted to obtain a prediction result of an anxiety mileage value through the target prediction model.
Drawings
Fig. 1 is a flow chart of a method for predicting an anxiety mileage value according to an embodiment of the present invention;
fig. 2 is a block diagram of an anxiety mileage predicting device according to an embodiment of the present invention.
Description of the embodiments
The principles and features of the present invention are described below with reference to the drawings, the examples are illustrated for the purpose of illustrating the invention and are not to be construed as limiting the scope of the invention.
Fig. 1 is a flow chart of a method for predicting an anxiety mileage value according to an embodiment of the present invention.
As shown in fig. 1, a method for predicting an anxiety mileage value includes the steps of:
s1: importing a plurality of original electric vehicle running data, and cleaning all the original electric vehicle running data to obtain a plurality of cleaned running data;
s2: screening and analyzing all the cleaned driving data to obtain a mileage anxiety data set;
s3: dividing the mileage anxiety data set to obtain a mileage anxiety training set and a mileage anxiety testing set;
s4: building a training model, and training the training model according to the mileage anxiety training set to obtain an original prediction model;
s5: testing the original prediction model according to the mileage anxiety testing set to obtain a target prediction model;
S6: and importing the running data of the electric automobile to be predicted, and predicting the running data of the electric automobile to be predicted through the target prediction model to obtain a prediction result of the anxiety mileage value.
It should be understood that, data cleaning is performed on all original electric vehicle history running data (i.e., the original electric vehicle running data), data containing missing values is deleted, data with messy codes due to signal differences is deleted, and unreasonable data is discarded, so that a plurality of intermediate electric vehicle history running data (i.e., the cleaned running data) are finally obtained.
In particular, anxiety mileage influencing factors are numerous, and a nonlinear relationship is presented between these factors and anxiety mileage. When PSO-BP is used for solving the nonlinear regression problem, proper characteristic data are selected through researching characteristic factors, and the relationship between anxiety mileage and influence factors thereof is researched through data analysis. The improved PSO algorithm is used for optimizing the initial weight and the threshold of the PSO-BP neural network so as to improve the convergence speed of anxiety mileage prediction and avoid sinking into a local optimal value. And inputting a test set (namely the mileage anxiety test set) into the trained predictive network model (namely the original predictive model) to obtain an anxiety mileage predicted value and an evaluation index.
It should be understood that, according to the anxiety distance prediction optimization model (i.e., the target prediction model), prediction processing is performed on the data to be predicted (i.e., the driving data of the electric vehicle to be predicted), so as to obtain a prediction result of the anxiety distance data (i.e., a prediction result of the anxiety distance value).
In the embodiment, the data of the original electric vehicle driving data are cleaned to obtain the cleaned driving data, the cleaned driving data are screened and analyzed to obtain the mileage anxiety data set, the mileage anxiety data set is divided to obtain the mileage anxiety training set and the mileage anxiety testing set, the training model is trained according to the mileage anxiety training set to obtain the original prediction model, the target prediction model is tested according to the mileage anxiety testing set to obtain the prediction result of the anxiety mileage value, and the prediction result of the anxiety mileage value is obtained through the prediction of the target prediction model to the electric vehicle driving data.
Optionally, as an embodiment of the present invention, the post-cleaning driving data includes a driving distance, a driving distance position corresponding to the driving distance, and an electric vehicle operation parameter;
the process of S2 includes:
sequencing all the electric automobile operation parameters according to the sequence from small to large of the driving mileage to obtain a plurality of sequenced automobile operation parameters;
screening a plurality of ordered automobile operation parameters according to preset charging state information, and taking the driving mileage positions corresponding to the ordered automobile operation parameters obtained after screening as charging points so as to obtain a plurality of charging points;
screening a plurality of ordered automobile operation parameters according to preset automobile state information, and taking the driving mileage positions corresponding to the ordered automobile operation parameters obtained after screening as starting points so as to obtain a plurality of starting points;
taking the previous starting point of each charging point as an anxiety point, thereby obtaining a plurality of anxiety points;
respectively differencing the driving mileage corresponding to each charging point and the driving mileage corresponding to each anxiety point to obtain the anxiety mileage of each anxiety point;
taking the anxiety mileage of all the anxiety points and the electric vehicle operation parameters corresponding to all the anxiety points as a mileage anxiety data set to be processed;
And screening the anxiety mileage of the anxiety points and the electric vehicle operation parameters corresponding to the anxiety points in the mileage anxiety data set to be processed by using a Pearson correlation coefficient analysis method to obtain a mileage anxiety data set.
It will be appreciated that the target anxiety mileage data (i.e., the pending mileage anxiety data set) is feature extracted by spearman correlation analysis and the target anxiety mileage data set (i.e., the mileage anxiety data set) is constructed by collecting the appropriate feature data.
Specifically, the obtained historical driving data (i.e. the driving data after cleaning) of the middle electric automobile is utilized to mark the charging point of the electric automobile according to the change of the charging state data (i.e. the preset charging state information), mark the starting point of the electric automobile according to the change of the automobile state data and the like (i.e. the preset automobile state information), mark the starting point before the charging point as an anxiety point, calculate the driving range from the anxiety point to the charging point as an anxiety mileage, and finally record the data (i.e. the operation parameters of the electric automobile) when the electric automobile is in the anxiety point and the corresponding anxiety mileage are manufactured as a target anxiety mileage original data set (i.e. the to-be-processed mileage anxiety data set).
It should be understood that, the spearman correlation coefficient analysis is performed on the collected characteristic data (i.e. the mileage anxiety data set to be processed), the characteristic data with the proper correlation coefficient is selected to be made into a data set, and 4524 data (i.e. the mileage anxiety data set) are made.
In the embodiment, the mileage anxiety data set is obtained by screening and analyzing all the cleaned driving data, effective data can be extracted, and data with proper correlation coefficient is selected, so that the anxiety mileage of the electric automobile driver can be accurately predicted, and the problem of insufficient prediction of the anxiety mileage value of the existing electric automobile driver is solved.
Optionally, as an embodiment of the present invention, the process of S4 includes:
s41: constructing a PSO-BP neural network model;
s42: obtaining a plurality of primary particle positions and a plurality of primary initial velocities corresponding to the primary particle positions from the PSO-BP neural network model, and respectively carrying out initialization processing on each primary particle position and the primary initial velocity corresponding to each primary particle position to correspondingly obtain an initialized particle position of each primary particle position and an initialized initial velocity corresponding to each primary particle position;
S43: parameter updating is carried out on the PSO-BP neural network model according to the initialized particle positions of the original particle positions, and updated models of the original particle positions are correspondingly obtained;
s44: training the updated model of each primary particle position according to the mileage anxiety training set to obtain a plurality of anxiety mileage predicted values of each primary particle position;
s45: importing actual anxiety mileage values corresponding to the predicted anxiety mileage values, and calculating initial fitness values according to the predicted anxiety mileage values and the actual anxiety mileage values of the original particle positions to obtain the initial fitness values of the original particle positions;
s46: and analyzing the target fitness value according to all the original particle positions, all the initial fitness values of the original particle positions and all the initial initialization speeds corresponding to the original particle positions to obtain the target fitness value, and taking an updated model corresponding to the target fitness value as an original prediction model.
It should be appreciated that a training model is constructed and a predicted optimization model of anxiety mileage (i.e., the original predictive model) is derived from the anxiety mileage dataset (i.e., the mileage anxiety training set) using a modified PSO-BP algorithm.
It should be appreciated that a PSO-BP neural network model is constructed.
It should be appreciated that the PSO-BP training model (i.e., the PSO-BP neural network model) is utilized to train a training set (i.e., the mileage anxiety training set) based on the supported improved PSO-BP algorithm. A series of parameters such as particle count, maximum iteration number, inertial weight, learning factor and the like are set.
Specifically, the dimension sets: establishing a mapping relation between a BP neural network and a particle swarm algorithm: establishing a particle swarm search space with the same weight as the neural network and the same threshold number, and enabling the weight, the threshold value and the dimension value of the particles in the search space to be in one-to-one correspondence, wherein the calculation formula of the dimension of the search space is as follows:
Figure SMS_1
in the formula ,
Figure SMS_2
is a search space dimension; />
Figure SMS_3
The number of weights from the input layer to the hidden layer is represented; />
Figure SMS_4
Representation->
Figure SMS_5
A hidden layer node; />
Figure SMS_6
The number of weights from the hidden layer to the output layer is represented; />
Figure SMS_7
Indicating the number of neurons in the output layer. Initializing positions (namely the original particle positions) and initial velocities (namely the original initial velocities) of all particles, calculating the fitness (namely the initial fitness value) of the particles at the moment, recording the positions of the particles at the moment, and searching for individual optimal and global optimal values of the current particles.
In the embodiment, the training model is trained according to the mileage anxiety training set to obtain the original prediction model, and compared with the existing anxiety mileage value prediction method, the method provided by the invention overcomes the defects that the traditional anxiety mileage value prediction method is large in workload and easy to fall into local optimum and has poor prediction recognition on the electric automobile driver, and can accurately predict the anxiety mileage of the electric automobile driver.
Optionally, as an embodiment of the present invention, the process of S45 includes:
importing actual anxiety mileage values corresponding to the respective predicted anxiety mileage values;
based on a first formula, calculating an initial fitness value according to a plurality of predicted anxiety mileage values and actual anxiety mileage values of each original particle position to obtain the initial fitness value of each original particle position, wherein the first formula is as follows:
Figure SMS_8
，
wherein ,
Figure SMS_9
for the initial fitness value, +.>
Figure SMS_10
For the number of anxiety mileage predictors, +.>
Figure SMS_11
Is->
Figure SMS_12
Actual anxiety mileage values->
Figure SMS_13
Is->
Figure SMS_14
And a predicted anxiety distance value.
It should be understood that the fitness function is calculated as follows:
Figure SMS_15
wherein N is the number of samples;
Figure SMS_16
for actual anxiety mileage value +.>
Figure SMS_17
Is the predicted value of anxiety mileage.
In the embodiment, the initial fitness value is obtained by calculating the initial fitness value based on the first formula according to the anxiety mileage predicted value and the anxiety mileage actual value, so that a foundation is laid for subsequent data processing, the anxiety mileage of the electric automobile driver can be accurately predicted, and the problem of insufficient prediction of the anxiety mileage value of the existing electric automobile driver is solved.
Optionally, as an embodiment of the present invention, the process of S46 includes:
s461: obtaining the current iteration times, screening the minimum value of the initial fitness values of all the original particle positions in all the iteration times, obtaining a global optimal fitness value after screening, and taking the original particle position corresponding to the global optimal fitness value as a global optimal particle position;
s462: screening the minimum value of the initial fitness value of each primary particle position in all iteration times, correspondingly obtaining the individual optimal fitness value of each primary particle position after screening, taking the initial velocity corresponding to the individual optimal fitness value of the primary particle position as the individual optimal initial velocity of the primary particle position, and taking the primary particle position corresponding to the individual optimal fitness value of the primary particle position as the individual optimal particle position of the primary particle position;
S463: judging whether the current iteration times are equal to preset times, if not, executing S464-S466; if yes, then S467 is performed;
s464: updating the particle positions according to the global optimal fitness value, the global optimal particle positions, the individual optimal fitness value of each original particle position, the individual optimal particle positions and the individual optimal initial speed to obtain updated particle positions of each original particle position;
s465: updating the initial velocity according to the updated particle positions of the original particle positions and the individual optimal initial velocity to obtain updated initial velocities of the original particle positions;
s466: taking the updated particle position of the original particle position as an initialized particle position of the original particle position of the next iteration number, taking the updated initial velocity of the original particle position as an initialized initial velocity of the original particle position of the next iteration number, and returning to S43;
s467; and taking the global optimal fitness value as a target fitness value, and taking an updated model corresponding to the target fitness value as an original prediction model.
It will be appreciated that the fitness of the particles (i.e. the initial fitness value) is compared with an individual optimum (i.e. the individual optimum fitness value) and a global optimum (i.e. the global optimum fitness value). If the fitness of the particles (i.e. the initial fitness value) is smaller than the individual optimum (i.e. the individual optimum fitness value), the current value is taken as the individual optimum (i.e. the individual optimum fitness value); if the current global optimum (i.e. the global optimum) is greater than the individual optimum (i.e. the individual optimum) of the particle swarm, the global optimum (i.e. the global optimum) of the particle swarm is the individual optimum (i.e. the individual optimum) at that time.
Specifically, the speed and the position of the particles are updated, an optimal individual obtained by a PSO algorithm through the updated position (i.e. the updated particle position) and the speed (i.e. the updated initial speed) is used as an initial weight and a threshold value of the BP neural network to train the BP neural network, when the network training reaches the maximum iteration number (i.e. the preset number), the training is ended, then the result is output, otherwise, the position and the speed of the particles are updated, and iteration is continued until the training is ended.
In the embodiment, the target fitness value is obtained by analyzing the target fitness value according to the original particle position, the initial fitness value and the initial initialization speed, and the updated model corresponding to the target fitness value is used as the original prediction model.
Optionally, as an embodiment of the present invention, the process of S464 includes:
based on a second formula, updating the particle positions according to the global optimal fitness value, the global optimal particle positions, the individual optimal fitness value of each original particle position, the individual optimal particle positions and the individual optimal initial speed to obtain updated particle positions of each original particle position, wherein the second formula is as follows:
Figure SMS_18
，/>
wherein ,
Figure SMS_19
，/>
Figure SMS_20
，
wherein ,
Figure SMS_21
，
wherein ,
Figure SMS_43
is->
Figure SMS_46
First iteration number->
Figure SMS_49
Updated particle positions of the individual primary particle positions, respectively>
Figure SMS_23
For contraction factor, ++>
Figure SMS_31
For the current iteration number>
Figure SMS_35
Is->
Figure SMS_40
Inertia weight of number of iterations, +.>
Figure SMS_28
Is->
Figure SMS_32
First iteration number->
Figure SMS_34
Individual optimal initial velocity of individual primary particle positions, < >>
Figure SMS_37
and />
Figure SMS_39
Are all acceleration factors, and ∈ ->
Figure SMS_42
>0，/>
Figure SMS_45
>0，/>
Figure SMS_48
and />
Figure SMS_41
All have the value range of 0,1]Random function of->
Figure SMS_44
Is->
Figure SMS_47
First iteration number->
Figure SMS_50
Individual optimum fitness value of individual primary particle positions,/->
Figure SMS_24
Is->
Figure SMS_27
First iteration number->
Figure SMS_29
Individual optimum particle positions of the individual primary particle positions, < >>
Figure SMS_33
For global optimum fitness value, +.>
Figure SMS_22
For global optimal particle position,/->
Figure SMS_30
For collectingShrink coefficient->
Figure SMS_36
For total acceleration factor, ++>
Figure SMS_38
For the maximum value of the inertial weight, +.>
Figure SMS_25
Is the minimum value of inertial weight, +.>
Figure SMS_26
Is a preset number of times.
Specifically, parameters in a neural network model are updated using a PSO algorithm, wherein a population X containing n particles exists in the model, wherein
Figure SMS_51
. Let the corresponding vector of the ith particle in the population in D-dimensional space be +.>
Figure SMS_52
The vector is the position of the ith particle in D-dimensional space, and the particle position is calculated according to the objective function >
Figure SMS_53
Is used for the adaptation value of the (c). Let the vector corresponding to the speed of the ith particle be +.>
Figure SMS_54
The vector of the individual extremum of the particle can be expressed as +.>
Figure SMS_55
The vector of population X population extremum can be expressed as +.>
Figure SMS_56
. The individual extremum and the population extremum calculate the speed and position of the updated particles by the following formula:
Figure SMS_57
Figure SMS_58
in the formula ,
Figure SMS_59
is the current speed; d=1, 2, l, d; i=1, 2, l, n; k is the iteration number; />
Figure SMS_60
、/>
Figure SMS_61
Represents an acceleration factor, and->
Figure SMS_62
>0，/>
Figure SMS_63
>0；/>
Figure SMS_64
、/>
Figure SMS_65
Indicating a value range of 0,1]A random function.
The particle swarm algorithm is further improved, the search proportion of local optimum and global optimum is better determined, parameters are faster and better updated, inertial weights are introduced into the basic PSO algorithm to improve the search performance of the particle swarm, and an update formula of the particle speed after the inertial weights are introduced is as follows:
Figure SMS_66
Figure SMS_67
in the formula ,
Figure SMS_68
is inertial weight, ++>
Figure SMS_69
Is the maximum velocity of the particle swarm when flying in d-dimensional space.
The use of inertial weights can significantly improve the convergence rate of the algorithm.
Figure SMS_70
The calculation formula of (2) is as follows:
Figure SMS_71
in the formula ,
Figure SMS_72
for maximum value of inertial weight, +.>
Figure SMS_73
K is the current iteration number, and gen is the total iteration number, which is the minimum value of the inertia weight.
The particle swarm algorithm is further improved, the limit of the particle speed boundary is eliminated, the convergence of the PSO algorithm is quickened, parameters are updated faster and better, a shrinkage factor is introduced into the basic PSO algorithm, and the shrinkage factor can change the acceleration factor
Figure SMS_74
and />
Figure SMS_75
Thereby changing the velocity of the particles. The updated formula for particle velocity after the introduction of the shrink factor is: />
Figure SMS_76
Figure SMS_77
in the formula ,
Figure SMS_78
for contraction factor, ++>
Figure SMS_79
For total acceleration factor, ++>
Figure SMS_80
>4, K is the shrinkage factor.
The improved PSO-BP neural network model is constructed, the specific implementation process of the improved PSO algorithm is to integrate the inertial weight and the contraction factor, the calculation of the inertial weight adopts the calculation formula of the inertial weight, the calculation of the contraction factor adopts the update formula of the particle speed after the contraction factor is introduced, and the speed update formula of the improved PSO algorithm is obtained after the inertial weight and the contraction factor are introduced as follows:
Figure SMS_81
the specific implementation process of the improved PSO algorithm is to integrate the inertial weight and the contraction factor, the calculation of the inertial weight adopts the formula, the calculation of the contraction factor adopts the formula, and the speed update formula of the improved PSO algorithm obtained after the inertial weight and the contraction factor are introduced is as follows:
Figure SMS_82
。
in the above embodiment, the updated particles are obtained based on the second formula according to the global optimal fitness value, the global optimal particle position, the individual optimal fitness value, the individual optimal particle position and the particle position of the individual optimal initial velocity, so that the searching performance of the particle swarm and the convergence speed of the algorithm are improved, the limitation of the particle speed boundary is eliminated, the convergence of the PSO algorithm is accelerated, and the parameters are updated faster and better.
Optionally, as an embodiment of the present invention, the process of S465 includes:
based on a third formula, updating the initial velocity according to the updated particle positions of the original particle positions and the individual optimal initial velocity to obtain updated initial velocities of the original particle positions, wherein the third formula is as follows:
Figure SMS_83
，
wherein ,
Figure SMS_85
is->
Figure SMS_88
First iteration number->
Figure SMS_92
Updated initial velocity of the positions of the individual primary particles, < >>
Figure SMS_86
Is->
Figure SMS_87
First iteration number->
Figure SMS_90
Updated particle positions of the individual primary particle positions, respectively>
Figure SMS_91
Is->
Figure SMS_84
First iteration number->
Figure SMS_89
Individual optimum initial velocities for individual primary particle positions.
In the embodiment, based on the third formula, the updated initial speed is obtained by updating the initial speed according to the updated particle position and the individual optimal initial speed, and compared with the existing anxiety mileage value prediction method, the method overcomes the defects that the traditional anxiety mileage value prediction method is large in workload and easy to sink into local optimum and poor in prediction and identification of the electric automobile driver, and can accurately predict the anxiety mileage of the electric automobile driver.
Optionally, as another embodiment of the present invention, the present invention further includes:
according to the prediction result of anxiety mileage data, converting the anxiety mileage data into anxiety values, analyzing the anxiety values in the time and space dimensions of the city, and researching a proper charging pile addressing scheme, wherein the charging pile addressing scheme specifically comprises the following steps:
According to the prediction result of the anxiety mileage data, the anxiety value is converted, and the single trip mileage of the electric automobile in the city is mostly not more than one hundred kilometers, so the anxiety mileage and the anxiety value are converted as shown in the following formula:
Figure SMS_93
/>
in the formula ,
Figure SMS_94
anxiety value->
Figure SMS_95
Anxiety mileage predictors.
And (3) analyzing in the time and space dimensions of the city to develop a proper charging pile addressing scheme. Grid differentiation can be carried out on a certain city to obtain a plurality of grid areas, the identified anxiety electric vehicle driver carries out visualized numerical display on the grid area where the identified anxiety electric vehicle driver is located at a certain moment, finally, the historical and current regional anxiety values of the area can be obtained, and the important point area is deeply analyzed and researched in the time dimension.
It should be understood that, according to the prediction result of the anxiety mileage data, the anxiety value is converted, and analysis is performed in the time and space dimensions of the city, so as to develop a suitable charging pile addressing scheme. Grid differentiation can be carried out on a certain city to obtain a plurality of grid areas, the identified anxiety electric vehicle driver carries out visualized numerical display on the grid area where the identified anxiety electric vehicle driver is located at a certain moment, finally, the historical and current regional anxiety values of the area can be obtained, and the important point area is deeply analyzed and researched in the time dimension.
Alternatively, as another embodiment of the present invention, the present invention uses Root Mean Square Error (RMSE) and mean percentage error (MAPE) as the indexes for evaluating the performance of the model, and the calculation method thereof is shown in the following formula. The smaller the RMSE and MAPE values, the better the prediction performance of the model, and the first table is a model error comparison table.
Figure SMS_96
Figure SMS_97
TABLE 1
Figure SMS_98
As can be seen from the table, the invention is based on an improved BP neural network model, a particle swarm algorithm is introduced to obtain a PSO-BP neural network model, the current anxiety mileage of an electric automobile driver can be predicted, and then an inertia weight and a contraction factor are introduced to the PSO algorithm to obtain an improved PSO algorithm model. The anxiety mileage of the driver can intuitively reflect the mileage anxiety value of the current driver, and the anxiety mileage can be differentiated into the mileage anxiety value through a specific algorithm, and then the anxiety mileage is processed to serve the addressing work of the charging pile and the like. The invention predicts by using the improved PSO optimization algorithm model PSO-BP, and experimental results show that the accuracy of the prediction model is superior to that of the similar prediction model, and the prediction model can be directly applied to addressing of the charging piles, thereby having guiding and reference significance for planning and layout of the future optimized charging piles.
Alternatively, as another embodiment of the present invention, the present invention divides the mileage anxiety dataset into a training set and a testing set. And dividing the training set 8:2 ratio into a training set and a testing set, and finally forming an anxiety mileage training set, an anxiety mileage verification set and an anxiety mileage testing set.
Optionally, as another embodiment of the present invention, the present invention further includes:
all target mileage anxiety data are made into data sets, and the data sets are randomly divided into training sets, test sets, and verification sets.
Constructing a PSO-BP neural network model, and a fitness function, wherein the calculation formula is as follows:
Figure SMS_99
wherein N is the number of samples;
Figure SMS_100
for actual anxiety mileage value +.>
Figure SMS_101
Is the predicted value of anxiety mileage.
A series of parameters such as particle count, maximum iteration number, inertial weight, learning factor and the like are set. Initializing the positions and initial speeds of all particles, calculating the adaptability of the particles at the moment, recording the positions of the particles at the moment, and searching the individual optimal and global optimal values of the current particles;
the fitness of the particles is compared to an individual optimum and a global optimum. If the fitness of the particles is smaller than the individual optimum, taking the current value as the individual optimum value; if the current global optimum is greater than the individual optimum of the particle swarm, the global optimum of the particle swarm is the individual optimum at the moment;
updating parameters in a neural network model using a PSO algorithm, wherein a population X comprising n particles exists in the model, wherein
Figure SMS_102
. Let the corresponding vector of the ith particle in the population in D-dimensional space be
Figure SMS_103
The vector is the position of the ith particle in D-dimensional space, and the particle position is calculated according to the objective function>
Figure SMS_104
Is used for the adaptation value of the (c). Let the vector corresponding to the speed of the ith particle be +.>
Figure SMS_105
The vector of the individual extremum of the particle can be expressed as +.>
Figure SMS_106
The vector of population X population extremum can be expressed as +.>
Figure SMS_107
. The individual extremum and the population extremum calculate the speed and position of the updated particles by the following formula:
Figure SMS_108
Figure SMS_109
in the formula ,
Figure SMS_110
is the current speed; d=1, 2, l, d; i=1, 2, l, n; k is the iteration number; />
Figure SMS_111
、/>
Figure SMS_112
Represents an acceleration factor, and->
Figure SMS_113
>0，/>
Figure SMS_114
>0；/>
Figure SMS_115
、/>
Figure SMS_116
Indicating a value range of 0,1]A random function.
Training the BP neural network by taking an optimal individual obtained by updating the position and the speed by a PSO algorithm as an initial weight and a threshold value of the BP neural network, ending training when the network training reaches the maximum iteration number, outputting a result, otherwise updating the position and the speed of particles, and continuing iteration until the training is finished;
based on the PSO-BP algorithm, training a training set by using the training model constructed by the PSO-BP algorithm according to preset iterative training times.
Optionally, as another embodiment of the invention, the invention provides an anxiety mileage prediction model based on historical driving data of an electric vehicle driver by researching the influence of various factors on the mileage anxiety of the electric vehicle driver, recording the last mileage of the electric vehicle from starting to charging as anxiety mileage, and recording characteristic data in an initial period of a starting point to make a data set. The anxiety mileage is predicted by the improved anxiety mileage prediction model of PSO-BP, and the mileage anxiety can be quantized based on the anxiety mileage prediction model, so that the study of the addressing direction of the charging pile can be performed. And (3) carrying out anxiety mileage prediction on electric automobile users in a certain city at a certain moment, converting the anxiety mileage prediction into anxiety values, carrying out research and analysis on the time dimension and the space dimension of the city, and finally carrying out addressing construction of the charging piles.
Optionally, as another embodiment of the present invention, the present invention obtains historical driving data of the electric automobile from the electric automobile server, the electric automobile is marked as a charging point when being charged, and the charging point is marked as an anxiety point when being started last time. And recording the last mileage of the electric automobile from starting to charging as anxiety mileage, and recording data in the initial period of the anxiety point to obtain target anxiety mileage data. Data cleaning is carried out on all the original electric vehicle target anxiety mileage data to obtain a target anxiety mileage original data set; feature extraction is carried out on target automobile data through spearman correlation analysis, and feature points of a plurality of target automobile data are integrated to construct an anxiety mileage data set; building a training model, and training the training model according to a target automobile data set to obtain an anxiety mileage prediction model; optimizing the prediction model according to the target automobile data set to obtain an anxiety mileage prediction optimization model; according to the anxiety mileage prediction optimization model, predicting the anxiety mileage data to be predicted to obtain a prediction result of the anxiety mileage; it is converted into anxiety values, research analysis is performed on the time and space dimensions of the city, and finally research on addressing of charging piles is performed. The method for predicting the anxiety mileage value of the electric automobile driver provided by the invention can replace the traditional anxiety mileage value predicting method, and can be used for predicting and analyzing a large amount of electric automobile data in real time, so that the efficiency is high, the stability is strong, the accuracy is high, and the anxiety mileage value of the electric automobile driver can be predicted better.
Fig. 2 is a block diagram of an anxiety mileage predicting device according to an embodiment of the present invention.
Alternatively, as another embodiment of the present invention, as shown in fig. 2, an anxiety mileage value prediction device includes:
the data cleaning module is used for importing a plurality of original electric vehicle running data, and cleaning the data of all the original electric vehicle running data to obtain a plurality of cleaned running data;
the screening analysis module is used for screening and analyzing all the cleaned driving data to obtain a mileage anxiety data set;
the division module is used for dividing the mileage anxiety data set to obtain a mileage anxiety training set and a mileage anxiety testing set;
the model training module is used for constructing a training model, and training the training model according to the mileage anxiety training set to obtain an original prediction model;
the model test module is used for testing the original prediction model according to the mileage anxiety test set to obtain a target prediction model;
the prediction result obtaining module is used for importing the running data of the electric automobile to be predicted, and predicting the running data of the electric automobile to be predicted through the target prediction model to obtain a prediction result of the anxiety mileage value.
Alternatively, another embodiment of the present invention provides an anxiety mileage value prediction system including a memory, a processor, and a computer program stored in the memory and executable on the processor, which when executed by the processor, implements the anxiety mileage value prediction method as described above. The system may be a computer or the like.
Alternatively, another embodiment of the present invention provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the anxiety mileage value prediction method as described above.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the apparatus and units described above may refer to corresponding procedures in the foregoing method embodiments, which are not described herein again.
In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of elements is merely a logical functional division, and there may be additional divisions of actual implementation, e.g., multiple elements or components may be combined or integrated into another system, or some features may be omitted, or not performed.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment of the present invention.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention is essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.

Claims (10)

1. A method for predicting anxiety mileage values, comprising the steps of:
s1: importing a plurality of original electric vehicle running data, and cleaning all the original electric vehicle running data to obtain a plurality of cleaned running data;
s2: screening and analyzing all the cleaned driving data to obtain a mileage anxiety data set;
s3: dividing the mileage anxiety data set to obtain a mileage anxiety training set and a mileage anxiety testing set;
s4: building a training model, and training the training model according to the mileage anxiety training set to obtain an original prediction model;
s5: testing the original prediction model according to the mileage anxiety testing set to obtain a target prediction model;
s6: and importing the running data of the electric automobile to be predicted, and predicting the running data of the electric automobile to be predicted through the target prediction model to obtain a prediction result of the anxiety mileage value.
2. The anxiety distance value prediction method according to claim 1, wherein said post-cleaning travel data includes a travel distance, a travel distance position corresponding to said travel distance, and an electric vehicle operation parameter;
The process of S2 includes:
sequencing all the electric automobile operation parameters according to the sequence from small to large of the driving mileage to obtain a plurality of sequenced automobile operation parameters;
screening a plurality of ordered automobile operation parameters according to preset charging state information, and taking the driving mileage positions corresponding to the ordered automobile operation parameters obtained after screening as charging points so as to obtain a plurality of charging points;
screening a plurality of ordered automobile operation parameters according to preset automobile state information, and taking the driving mileage positions corresponding to the ordered automobile operation parameters obtained after screening as starting points so as to obtain a plurality of starting points;
taking the previous starting point of each charging point as an anxiety point, thereby obtaining a plurality of anxiety points;
respectively differencing the driving mileage corresponding to each charging point and the driving mileage corresponding to each anxiety point to obtain the anxiety mileage of each anxiety point;
taking the anxiety mileage of all the anxiety points and the electric vehicle operation parameters corresponding to all the anxiety points as a mileage anxiety data set to be processed;
and screening the anxiety mileage of the anxiety points and the electric vehicle operation parameters corresponding to the anxiety points in the mileage anxiety data set to be processed by using a Pearson correlation coefficient analysis method to obtain a mileage anxiety data set.
3. The anxiety distance value prediction method according to claim 1, wherein said S4 process comprises:
s41: constructing a PSO-BP neural network model;
s42: obtaining a plurality of primary particle positions and a plurality of primary initial velocities corresponding to the primary particle positions from the PSO-BP neural network model, and respectively carrying out initialization processing on each primary particle position and the primary initial velocity corresponding to each primary particle position to correspondingly obtain an initialized particle position of each primary particle position and an initialized initial velocity corresponding to each primary particle position;
s43: parameter updating is carried out on the PSO-BP neural network model according to the initialized particle positions of the original particle positions, and updated models of the original particle positions are correspondingly obtained;
s44: training the updated model of each primary particle position according to the mileage anxiety training set to obtain a plurality of anxiety mileage predicted values of each primary particle position;
s45: importing actual anxiety mileage values corresponding to the predicted anxiety mileage values, and calculating initial fitness values according to the predicted anxiety mileage values and the actual anxiety mileage values of the original particle positions to obtain the initial fitness values of the original particle positions;
S46: and analyzing the target fitness value according to all the original particle positions, all the initial fitness values of the original particle positions and all the initial initialization speeds corresponding to the original particle positions to obtain the target fitness value, and taking an updated model corresponding to the target fitness value as an original prediction model.
4. A method of predicting an anxiety mileage value as set forth in claim 3, wherein the process of S45 includes:
importing actual anxiety mileage values corresponding to the respective predicted anxiety mileage values;
based on a first formula, calculating an initial fitness value according to a plurality of predicted anxiety mileage values and actual anxiety mileage values of each original particle position to obtain the initial fitness value of each original particle position, wherein the first formula is as follows:
Figure QLYQS_1
，
wherein ,
Figure QLYQS_2
for the initial fitness value, +.>
Figure QLYQS_3
For the number of anxiety mileage predictors, +.>
Figure QLYQS_4
Is->
Figure QLYQS_5
Actual anxiety mileage values->
Figure QLYQS_6
Is->
Figure QLYQS_7
And a predicted anxiety distance value.
5. A method of predicting an anxiety mileage value as set forth in claim 3, wherein the process of S46 includes:
s461: obtaining the current iteration times, screening the minimum value of the initial fitness values of all the original particle positions in all the iteration times, obtaining a global optimal fitness value after screening, and taking the original particle position corresponding to the global optimal fitness value as a global optimal particle position;
S462: screening the minimum value of the initial fitness value of each primary particle position in all iteration times, correspondingly obtaining the individual optimal fitness value of each primary particle position after screening, taking the initial velocity corresponding to the individual optimal fitness value of the primary particle position as the individual optimal initial velocity of the primary particle position, and taking the primary particle position corresponding to the individual optimal fitness value of the primary particle position as the individual optimal particle position of the primary particle position;
s463: judging whether the current iteration times are equal to preset times, if not, executing S464-S466; if yes, then S467 is performed;
s464: updating the particle positions according to the global optimal fitness value, the global optimal particle positions, the individual optimal fitness value of each original particle position, the individual optimal particle positions and the individual optimal initial speed to obtain updated particle positions of each original particle position;
s465: updating the initial velocity according to the updated particle positions of the original particle positions and the individual optimal initial velocity to obtain updated initial velocities of the original particle positions;
S466: taking the updated particle position of the original particle position as an initialized particle position of the original particle position of the next iteration number, taking the updated initial velocity of the original particle position as an initialized initial velocity of the original particle position of the next iteration number, and returning to S43;
s467; and taking the global optimal fitness value as a target fitness value, and taking an updated model corresponding to the target fitness value as an original prediction model.
6. The anxiety distance value prediction method according to claim 5, wherein said S464 process comprises:
based on a second formula, updating the particle positions according to the global optimal fitness value, the global optimal particle positions, the individual optimal fitness value of each original particle position, the individual optimal particle positions and the individual optimal initial speed to obtain updated particle positions of each original particle position, wherein the second formula is as follows:
Figure QLYQS_8
，/>
wherein ,
Figure QLYQS_9
，/>
Figure QLYQS_10
，
wherein ,
Figure QLYQS_11
，
wherein ,
Figure QLYQS_21
is->
Figure QLYQS_26
First iteration number->
Figure QLYQS_29
Updated particle positions of the individual primary particle positions, respectively>
Figure QLYQS_12
For contraction factor, ++>
Figure QLYQS_14
For the current iteration number>
Figure QLYQS_16
Is->
Figure QLYQS_18
Inertia weight of number of iterations, +. >
Figure QLYQS_22
Is->
Figure QLYQS_25
First iteration number->
Figure QLYQS_33
Individual optimal initial velocity of individual primary particle positions, < >>
Figure QLYQS_35
and />
Figure QLYQS_32
Are all acceleration factors, and ∈ ->
Figure QLYQS_37
>0，/>
Figure QLYQS_39
>0，/>
Figure QLYQS_40
and />
Figure QLYQS_24
All have the value range of 0,1]Random function of->
Figure QLYQS_27
Is->
Figure QLYQS_31
First iteration number->
Figure QLYQS_34
Individual optimum fitness values for individual primary particle positions,
Figure QLYQS_13
is->
Figure QLYQS_15
First iteration number->
Figure QLYQS_17
Individual optimum particle positions of the individual primary particle positions, < >>
Figure QLYQS_19
For global optimum fitness value, +.>
Figure QLYQS_20
For global optimal particle position,/->
Figure QLYQS_23
For the contraction coefficient->
Figure QLYQS_28
For total acceleration factor, ++>
Figure QLYQS_30
For the maximum value of the inertial weight, +.>
Figure QLYQS_36
Is the minimum value of inertial weight, +.>
Figure QLYQS_38
Is a preset number of times.
7. The anxiety distance value prediction method according to claim 5, wherein said step S465 comprises:
based on a third formula, updating the initial velocity according to the updated particle positions of the original particle positions and the individual optimal initial velocity to obtain updated initial velocities of the original particle positions, wherein the third formula is as follows:
Figure QLYQS_41
，
wherein ,
Figure QLYQS_42
is->
Figure QLYQS_46
First iteration number->
Figure QLYQS_50
Updated initial velocity of the positions of the individual primary particles, < >>
Figure QLYQS_44
Is->
Figure QLYQS_45
First iteration number->
Figure QLYQS_47
Updated particle positions of the individual primary particle positions, respectively>
Figure QLYQS_49
Is->
Figure QLYQS_43
First iteration number- >
Figure QLYQS_48
Individual optimum initial velocities for individual primary particle positions.
8. An anxiety mileage prediction device, comprising:
the data cleaning module is used for importing a plurality of original electric vehicle running data, and cleaning the data of all the original electric vehicle running data to obtain a plurality of cleaned running data;
the screening analysis module is used for screening and analyzing all the cleaned driving data to obtain a mileage anxiety data set;
the division module is used for dividing the mileage anxiety data set to obtain a mileage anxiety training set and a mileage anxiety testing set;
the model training module is used for constructing a training model, and training the training model according to the mileage anxiety training set to obtain an original prediction model;
the model test module is used for testing the original prediction model according to the mileage anxiety test set to obtain a target prediction model;
the prediction result obtaining module is used for importing the running data of the electric automobile to be predicted, and predicting the running data of the electric automobile to be predicted through the target prediction model to obtain a prediction result of the anxiety mileage value.
9. A anxiety distance value prediction system comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the anxiety distance value prediction method of any one of claims 1 to 7 is implemented when the computer program is executed by the processor.
10. A computer readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the anxiety mileage value prediction method according to any one of claims 1 to 7.
CN202310115798.3A 2023-02-15 2023-02-15 Anxiety mileage value prediction method, anxiety mileage value prediction device, anxiety mileage value prediction system and storage medium Pending CN116029329A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310115798.3A CN116029329A (en) 2023-02-15 2023-02-15 Anxiety mileage value prediction method, anxiety mileage value prediction device, anxiety mileage value prediction system and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310115798.3A CN116029329A (en) 2023-02-15 2023-02-15 Anxiety mileage value prediction method, anxiety mileage value prediction device, anxiety mileage value prediction system and storage medium

Publications (1)

Publication Number Publication Date
CN116029329A true CN116029329A (en) 2023-04-28

Family

ID=86091162

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310115798.3A Pending CN116029329A (en) 2023-02-15 2023-02-15 Anxiety mileage value prediction method, anxiety mileage value prediction device, anxiety mileage value prediction system and storage medium

Country Status (1)

Country Link
CN (1) CN116029329A (en)

Citations (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103729695A (en) * 2014-01-06 2014-04-16 国家电网公司 Short-term power load forecasting method based on particle swarm and BP neural network
US20140278104A1 (en) * 2013-03-13 2014-09-18 Ford Global Technologies, Llc Electric vehicle emergency recharge assistance
CN107139762A (en) * 2017-06-05 2017-09-08 吉林大学 A kind of electric automobile optimization charge control method and its system
CN107199903A (en) * 2017-05-31 2017-09-26 国网山东省电力公司莱芜供电公司 A kind of electric automobile accesses the intelligent charge strategy of power distribution network
JP2018133844A (en) * 2017-02-13 2018-08-23 株式会社日立製作所 Ev charge/discharge control device
CN108764634A (en) * 2018-04-24 2018-11-06 河海大学 A kind of electric automobile charging station dynamic programming method for considering charge requirement and increasing
CN109447233A (en) * 2018-11-09 2019-03-08 山东大学 Electric car charge and discharge dispatching method and system
CN110288212A (en) * 2019-06-14 2019-09-27 石家庄铁道大学 Electric taxi based on improved MOPSO creates charging station site selecting method
CN111523930A (en) * 2020-04-17 2020-08-11 国网上海市电力公司 Planning method for site selection and service range division of urban electric taxi charging station
US20210331601A1 (en) * 2020-04-23 2021-10-28 Yanshan University Dispatching method and system for electric vehicle battery swapping station
US20220188729A1 (en) * 2020-12-10 2022-06-16 Honda Motor Co., Ltd. System and method for placement optimization of public electric vehicle charging stations using telematics data
CN114707292A (en) * 2022-01-21 2022-07-05 东北电力大学 Voltage stability analysis method for power distribution network containing electric automobile
CN114819370A (en) * 2022-05-09 2022-07-29 国网浙江省电力有限公司 Improved hybrid algorithm-based electric vehicle charging station location and volume selection method and device
CN114880806A (en) * 2022-05-25 2022-08-09 国网天津市电力公司 New energy automobile sales prediction model parameter optimization method based on particle swarm optimization
US20220283240A1 (en) * 2021-03-04 2022-09-08 Wuhan University Method and system for estimating state of health of battery pack
CN115544900A (en) * 2022-11-28 2022-12-30 深圳联友科技有限公司 Method for analyzing influence factors of endurance mileage of electric vehicle based on Shap algorithm
CN115656840A (en) * 2022-12-27 2023-01-31 武汉工程大学 Method, device, system and storage medium for predicting battery charging remaining time

Patent Citations (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140278104A1 (en) * 2013-03-13 2014-09-18 Ford Global Technologies, Llc Electric vehicle emergency recharge assistance
CN103729695A (en) * 2014-01-06 2014-04-16 国家电网公司 Short-term power load forecasting method based on particle swarm and BP neural network
JP2018133844A (en) * 2017-02-13 2018-08-23 株式会社日立製作所 Ev charge/discharge control device
CN107199903A (en) * 2017-05-31 2017-09-26 国网山东省电力公司莱芜供电公司 A kind of electric automobile accesses the intelligent charge strategy of power distribution network
CN107139762A (en) * 2017-06-05 2017-09-08 吉林大学 A kind of electric automobile optimization charge control method and its system
CN108764634A (en) * 2018-04-24 2018-11-06 河海大学 A kind of electric automobile charging station dynamic programming method for considering charge requirement and increasing
CN109447233A (en) * 2018-11-09 2019-03-08 山东大学 Electric car charge and discharge dispatching method and system
CN110288212A (en) * 2019-06-14 2019-09-27 石家庄铁道大学 Electric taxi based on improved MOPSO creates charging station site selecting method
CN111523930A (en) * 2020-04-17 2020-08-11 国网上海市电力公司 Planning method for site selection and service range division of urban electric taxi charging station
US20210331601A1 (en) * 2020-04-23 2021-10-28 Yanshan University Dispatching method and system for electric vehicle battery swapping station
US20220188729A1 (en) * 2020-12-10 2022-06-16 Honda Motor Co., Ltd. System and method for placement optimization of public electric vehicle charging stations using telematics data
US20220283240A1 (en) * 2021-03-04 2022-09-08 Wuhan University Method and system for estimating state of health of battery pack
CN114707292A (en) * 2022-01-21 2022-07-05 东北电力大学 Voltage stability analysis method for power distribution network containing electric automobile
CN114819370A (en) * 2022-05-09 2022-07-29 国网浙江省电力有限公司 Improved hybrid algorithm-based electric vehicle charging station location and volume selection method and device
CN114880806A (en) * 2022-05-25 2022-08-09 国网天津市电力公司 New energy automobile sales prediction model parameter optimization method based on particle swarm optimization
CN115544900A (en) * 2022-11-28 2022-12-30 深圳联友科技有限公司 Method for analyzing influence factors of endurance mileage of electric vehicle based on Shap algorithm
CN115656840A (en) * 2022-12-27 2023-01-31 武汉工程大学 Method, device, system and storage medium for predicting battery charging remaining time

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
HAN LIU ET AL.: "Research and Application of Range Anxiety in Electric Vehicle Drivers", 《PROCEEDINGS OF THE 3RD ASIA-PACI CONFERENCE ON IMAGE PROCESSING，ELECTRONICS AND COMPUTERS》, pages 525 - 529 *
张洪涛 等: "改进PSO优化LSSVM的锂电池SOC估算模型", 《华侨大学学报》, vol. 42, no. 2, pages 245 - 250 *
蔡主希: "《智能风控与反欺诈 体系算法与实践》", 机械工业出版社, pages: 27 - 28 *

Similar Documents

Publication Publication Date Title
Bi et al. Estimating remaining driving range of battery electric vehicles based on real-world data: A case study of Beijing, China
CN107909179B (en) Method for constructing prediction model of running condition of plug-in hybrid vehicle and vehicle energy management method
CN109214581B (en) Railway line wind speed prediction method considering wind direction and confidence interval
CN105849998A (en) Electricity-demand prediction device, electricity supply system, electricity-demand prediction method, and program
CN108556682B (en) Driving range prediction method, device and equipment
CN104239496A (en) Collaborative filtering method based on integration of fuzzy weight similarity measurement and clustering
CN113780665B (en) Private car stay position prediction method and system based on enhanced recurrent neural network
CN116572799B (en) Power battery charge duration prediction method, system and terminal based on deep learning
CN106777169A (en) A kind of user&#39;s trip hobby analysis method based on car networking data
CN112419711B (en) Closed parking lot parking demand prediction method based on improved GMDH algorithm
CN114511154A (en) Traffic flow prediction method based on space-time complex graph convolution network
CN115871450A (en) New energy automobile intelligent control method and system based on Internet of things
CN108830414B (en) Load prediction method for commercial charging area of electric automobile
CN115759462A (en) Charging behavior prediction method and device for electric vehicle user and electronic equipment
CN117400785A (en) Prediction method and device for pure electric endurance mileage, electronic equipment and vehicle
CN113642768A (en) Vehicle running energy consumption prediction method based on working condition reconstruction
CN116125289A (en) Lithium battery health state estimation method based on local voltage segment and ensemble learning
CN112098869B (en) Self-adaptive electric vehicle SOC estimation method based on big data
CN117390867A (en) Energy consumption evaluation method and device, electronic equipment and medium
CN116029329A (en) Anxiety mileage value prediction method, anxiety mileage value prediction device, anxiety mileage value prediction system and storage medium
CN112035536A (en) Electric automobile energy consumption prediction method considering dynamic road network traffic flow
CN114492169B (en) Electric automobile user portraying method and system considering user market response potential
CN115691140B (en) Analysis and prediction method for space-time distribution of automobile charging demand
CN116662860A (en) User portrait and classification method based on energy big data
CN116224078A (en) New energy automobile power battery internal resistance calculation and prediction method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20230428

RJ01 Rejection of invention patent application after publication